Beamforming design for two-antenna MISO-NOMA system with statistical CSI

In this paper, we study the problem of optimizing statistical beamforming design for each user in a two-antenna downlink multiple-input single-output (MISO) non-orthogonal multiple access (NOMA) system. The transmitter only possesses statistical information in the form of covariance matrices for each user’s link. The statistical beamforming designs are derived in maximizing the ergodic sum rate considering both low and high signal-to-interference-plus-noise ratio (SINR) extreme scenarios. In addition to the analytical investigations, we also conduct Monte-Carlo simulations in comparison and to affirm the accuracy of the derived ergodic sum rate expressions. Results show that the ergodic sum rate increases with the total transmit power and decreases with the power coefficient of near users. Comparing the optimized MISO-NOMA system with conventional MISO-OMA scheme, it is demonstrated that MISO-NOMA can significantly improve the ergodic sum rate.